Diet is a critical environmental variable that exerts a powerful influence on health and disease. Cancer, cardiovascular disease, obesity and type-2 diabetes are particularly sensitive to diet. There is a parallel rise in the incidence of these conditions and the consumption of foods that contribute to their development. While genetics and diet each unquestionably affect disease incidence, it is increasingly clear that the dynamic between genetics and diet is what determines health or disease status. Neither variable acts entirely independently of the other. Furthermore, the genetic influences may be both quantitative - affecting frequency of tumor development, and complex- involving multiple gene-gene interactions. It is possible to use the genetic diversity of different inbred strains of mice as the basis of a systematic, genome-wide hunt for loci that control if and how diet affects colorectal cancer. Our preliminary data show that when the A/J and C57BL/6 genomes are combined in F1 mice, and the carcinogen azoxymethane (AOM) is used to induce colorectal cancer, the tumor number in those F1 animals depends on diet. In the inbred parents, no diet responsive phenotype was seen. This demonstrates that A/J and C57BL/6 harbor interacting alleles that can combine to produce a diet-responsive, complex, quantitative affect on tumor number. In a narrowly-focused single Aim, I propose to use recombinant inbred (Rl) and chromosome substitution strains (CSS) derived from A/J and C57BL/6 as tools in the next incremental step to map genetic loci controling if and how diet affects colorectal cancer. I propose to maintain the BxA Rl panel and B.A CSS lines derived from this strain combination on either a control or "western style" diet that is high in fat, low in fiber, Ca+2, vitamin D3, choline and folic acid. After two weeks on the diet, mice will be given eight weekly injections of AOM to induce colon tumors and will be maintained on their respective diet for the duration of the experiment. Twenty-five weeks after the last AOM injection we will monitor tumor burden in mice and other qualitative tumor phenotypes. Using the publicly-available strain distribution patterns (SDP) known for the Rl and CSS mice, we will perform interval mapping to identify genomic regions controlling diet- responsive tumor phenotypes. This complex quantitative trait analysis will reveal loci at the heart of gene- nutrient interactions relevant to colon cancer and lays the foundation for subsequent work to identify the genes critical for those interactions. [unreadable] [unreadable] [unreadable]